The present invention relates to a technique for recognizing an object existing in a video in real time.
In the technical field described above, Patent Document 1 describes a technique for improving recognition speed by clustering characteristic amounts when recognizing a query image using a model dictionary generated from a model image in advance.
[Patent Document 1] Patent Publication JP-A-2011-22168